Using Linked Data to provide a different perspective on Software Architecture
As outlined in my previous post Linked Data and the SOA Software Development Process I am interested in using Linked Data to provided a more detailed view of SOA services.
A coupled of scenarios during the past week highlighted the value of the approach and also that it would benefit with extending the scope to include more information about the consumers of the SOA services and also the external data sources (in particular databases) used by the SOA services.
Both scenarios involved setting up environments for the development and testing of new functionality involving a number of different systems, with each system needing to be deployed at a specific version level.
The first scenario related to the software versions. The UML diagrams presented to describe the architecture were at too high a level to show the actual dependencies, but to add the level of detail needed would have made the diagrams too busy.
Although not yet complete the work already done to provide a Linked Data perspective of the SOA services enabled a more fined grained view of the actual dependencies. Knowing what the specific lower level dependencies were resulted in more flexibility with the actual deployment. In particular work could start on developing the new functionality for one component since it was not going to be affected by proposed changes in another component. On the original UML diagram both components were shown as requiring changes. The Linked Data perspective provided enough additional detail to see that the changes could happen in parallel.
The second scenario related to finding the owners of external data sources so that we could determine if they were available for use in a given test environment. Adding this ownership information to our Linked Data repository would speed up this part of the process in the future.